The need to improve the translation of risk between adverse events in non-human species and the clinical population was a topic addressed by several speakers at this year’s PepTalk: The Protein Science Week, one of the largest annual gatherings of protein science researchers in the world, organized by Cambridge Healthtech Institute, which took place on January 8-13, 2017 in San Diego, California.[a]
Researchers and professionals in the field of Antibody-drug Conjugates, or ADCs, discussed some of the issues that can cause difficultly predicting adverse events that occur in the clinic from pre-clinical data, and how they may be accounted for to develop better translation models.
Peripheral Neuropathy for vcMMAE ADCs
For current vcMMAE ADCs [b], incidences of peripheral neuropathy (PN) are high in the clinic, but this is often not observed in nonclinical toxicology studies.
Peripheral neuropathy, damage to the peripheral nerves, is generally the result of traumatic injuries, infections, metabolic problems, inherited causes and exposure to toxins. It causes weakness, numbness and pain, usually in hands and feet, but can also affect other areas of the body. Patients describe the pain cause as a result of peripheral neuropathy as stabbing, burning or tingling. Symptoms may improve, especially if PN is caused by a treatable condition. Medications can reduce the pain.
The lack of translatability from non-clinical to clinical studies presents many challenges. There is a need to better understand why standard nonclinical toxicology studies for IND regulation submissions have failed to accurately predict the occurrence of PN, and for new developments to be made in predictive models and screenings.
Moving ADCs into the Clinic
On the Wednesday, January 11th segment “Moving ADCs into the Clinic”, Nicola Stagg, Ph.D., Scientist and Toxicologist for Safety Assessment at Genentech, evaluated several hypothesis that could account for why animals do not show the high NP incidence that occurs in vcMMAE ADC treated humans, and clarified some of the many factors to consider. One of these is the challenge of accounting for differences between an animal model and clinical population, such as different testing techniques and dosing in animals, and prior treatments, drug-drug interactions, and pharmacogenetics in the clinical population, make translatability a challenge.
For instance, Stagg discussed how up to 86% of patients treated with vcMMAE ADCs receive prior chemotherapy that can cause PN, and several of these patients likely had co-morbidities that can cause the inheritance of neuropathy associated with PN, such as diabetes, alcoholism, nutritional deficiencies, and infectious diseases like AIDS and Lyme Disease. Additionally, clinical data suggests that more frequent dosing regimens are associated with increased PN, but there is no data to address this in nonclinical species.
To try to mitigate some of these issues, Stagg presented on how several key leanings can be implemented to improve PN predictability before clinical studies. For instance, exploring transgenic animal models with predisposing conditions to developing PN could provide a closer model to the some patients in the clinical populations.
There are also considerable pharmacokinetics factors, like differences in pharmacokinetic (PK) profile, duration of exposures, catabolite and metabolite excretion, and the distribution of peripheral nerves between species. In terms of duration of exosure, the amount of time that non-human species are studied may simply not be long enough to get an accurate PN prediction. PN occurance occurance is seen in humans between 22 and 32 weeks of being exposed to a vcMMAE ADC, while nonclinical toxicology studies are done for a maximum of 15 weeks. Stagg suggested that exposure should be done for a longer period of time to provide better comparable data. Additionally, Stagg suggested that incorporating an expanded neurohistopathology assessment o f peripheral nerves may also help improve modeling PN in the future.
While accounting for differences between animal models and clinical polualions are incredibly coplex, Stagg provided many suggestions on how to develop a predictive animal modeal that may provide atrategies to improve traenslation for PN in vcMMAE ADCs , and to screen next generation microtublule inhibitor ADCs for reduced PN.
Another presentation that focused on pre-clinical to clinical translation of adverse events was given by Shangchuing Chen, Ph.D., Scientist in Clinical Pharmacology at Genentech on Wednesday, January 11th Analytical Strategies segment. Chen presented on the opportunity for building a quantitative pre-clinical/clinical model that was developed using PK and PD data to translate neutropenia risk for multiple vc-MMAE ADCs.
Neutropenia, an abnormally low concentration of neutrophils in the blood, is a common adverse event in patients treated with a vc-MMAE ADC. It results in dose reduction, delay, and sometimes discontinuation of treatment. And while it is manageable, neutropenia has the potential to cause severe infection. Neutropenia is commonly detected in non-human species toxicology studies, but there is a need to build and validate a translational model that can better predict neutropenia risk for future ADCs before they reach the clinical population.
Chen explained Genentech’s mechanistic nuetropenial translational model and it’s parameters. This model takes into account species specific parameters like baseline neutrophil cell count, maturation time, and rebound of neutrophil cell as well as drug-specific parameters like the drug’s effect on progenitor cells. One main question with such model is whether or not the species specific parameters are independent of the drug being tested. According to Chen, parameters from individual pharmacokinetic/pharmacodynamic (PK/PD) analysis of several drugs were comparable, meaning that parameters are not dependent on drug.
According to Chen, then the challenge of whether or not these cyno-specific drug parameter can be translated to those established for humans arises. Unlike the translation for cytotoxic drugs that require correction of protein binding and sensitivity, drug-specific parameters are approximately comparable between cyno and human populations.
According to the data presented by Genentech, the drug specific parameters were comparable, mostly within two fold, between human and cyno for four out of the five drugs that were compared. This type of translational neutropenia model could help project human neutrophil count over time, which can help in the development o clinical risk mitigation strategies, and leading candidate selection of treatment with vcMMAE ADCs.
[a] PepTalk: The Protein Science Week is one of the largest annual gatherings of protein science researchers in the world. In its 16th year, PepTalk attracts nearly 1,300 experts from academia, biotech and pharma who come together for one week of intensive learning and networking to discover new opportunities and promising partnerships. This event covers a wide spectrum, from upstream protein R&D science to downstream biologics.
[b] vcMMAE ADC is an antibody-drug conjugate with potent antitumor activity by using the antimitotic agent, monomethyl auristatin E (MMAE), linked via the lysosomally cleavable dipeptide, valine-citrulline (vc) linker.